Alzheimer's disease is a progressive neurodegenerative disorder that remains challenging to predict due to its multifactorial etiology and the complexity of multimodal clinical data. Accurate forecasting of clinically relevant biomarkers, including diagnostic and quantitative measures, is essential for effective monitoring of disease progression. This work introduces L2C-TabPFN, a method that integrates a longitudinal-to-cross-sectional (L2C) transformation with a pre-trained Tabular Foundation Model (TabPFN) to predict Alzheimer's disease outcomes using the TADPOLE dataset. L2C-TabPFN converts sequential patient records into fixed-length feature vectors, enabling robust prediction of diagnosis, cognitive scores, and ventricular volume. Experimental results demonstrate that, while L2C-TabPFN achieves competitive performance on diagnostic and cognitive outcomes, it provides state-of-the-art results in ventricular volume prediction. This key imaging biomarker reflects neurodegeneration and progression in Alzheimer's disease. These findings highlight the potential of tabular foundational models for advancing longitudinal prediction of clinically relevant imaging markers in Alzheimer's disease.

W e address the problem of few-shot pattern detection, which aims to detect all instances of a given pattern, typically represented by a few exemplars, from an input image. Although similar problems have been studied in few-shot object counting and detection (FSCD), previous methods and their benchmarks have narrowed patterns of interest to object categories and often fail to localize non-object patterns. In this work, we propose a simple yet effective detector based on template matching and regression, dubbed TMR. While previous FSCD methods typically represent target exemplars as spatially collapsed prototypes and lose structural information, we revisit classic template matching and regression. It effectively preserves and leverages the spatial layout of exemplars through a minimalistic structure with a small number of learnable convolutional or projection layers on top of a frozen backbone. W e also introduce a new dataset, dubbed RPINE, which covers a wider range of patterns than existing object-centric datasets. Our method outperforms the state-of-the-art methods on the three benchmarks, RPINE, FSCD-147, and FSCD-LVIS, and demonstrates strong generalization in cross-dataset evaluation.

--In industrial big data scenarios, high-dimensional sparse matrices (HDI) are widely used to characterize high-order interaction relationships among massive nodes. The stochastic gradient descent-based latent factor analysis (SGD-LFA) method can effectively extract deep feature information embedded in HDI matrices. However, existing SGD-LFA methods exhibit significant limitations: their parameter update process relies solely on the instantaneous gradient information of current samples, failing to incorporate accumulated experiential knowledge from historical iterations or account for intrinsic correlations between samples, resulting in slow convergence speed and suboptimal generalization performance. Thus, this paper proposes a PILF model by developing a PI-accelerated SGD algorithm by integrating correlated instances and refining learning errors through proportional-integral (PI) control mechanism that current and historical information; Comparative experiments demonstrate the superior representation capability of the PILF model on HDI matrices. ATA serves as the foundation for industrial application development.

Graph homophily has been considered an essential property for message-passing neural networks (MPNN) in node classification. Recent findings suggest that performance is more closely tied to the consistency of neighborhood class distributions. We demonstrate that the MPNN performance depends on the number of components of the overall neighborhood distribution within a class. By breaking down the classes into their neighborhood distribution components, we increase measures of neighborhood distribution informativeness but do not observe an improvement in MPNN performance. We propose a Gumbel-Softmax-based rewiring method that reduces deviations in neighborhood distributions. Our results show that our new method enhances neighborhood informativeness, handles long-range dependencies, mitigates oversquashing, and increases the classification performance of the MPNN. The code is available at https://github.com/Bobowner/Gumbel-Softmax-MPNN.

Handling missing data in time series classification remains a significant challenge in various domains. Traditional methods often rely on imputation, which may introduce bias or fail to capture the underlying temporal dynamics. In this paper, we propose TANDEM (Temporal Attention-guided Neural Differential Equations for Missingness), an attention-guided neural differential equation framework that effectively classifies time series data with missing values. Our approach integrates raw observation, interpolated control path, and continuous latent dynamics through a novel attention mechanism, allowing the model to focus on the most informative aspects of the data. We evaluate TANDEM on 30 benchmark datasets and a real-world medical dataset, demonstrating its superiority over existing state-of-the-art methods. Our framework not only improves classification accuracy but also provides insights into the handling of missing data, making it a valuable tool in practice.

The synthesis of symbolic logic and deep learning has given rise to a new class of interpretable models that stand as a powerful alternative to opaque, black-box architectures. The pioneering work in this domain includes deep differentiable logic gate networks (LGNs) and their successor, convolutional LGNs, both proposed by Petersen et al. [1, 2]. These models establish a method for training networks of Boolean logic gates via gradient descent by creating differentiable relaxations of the logic operations. This core strategy yields transparent models from which logical rules can be extracted, while extensions like logical convolutions and pooling enable these architectures to scale effectively to complex classification tasks. The LGN paradigm was significantly generalized with the development of differentiable logic networks (DLNs) [3]. By removing prior constraints, such as fixed network topologies and binary-only inputs, DLNs broadened the framework's applicability to general tabular classification, resulting in highly efficient models that match or exceed the performance of traditional neural networks (NNs) on a wide range of benchmarks. The scope of these logic-based systems was expanded once more to handle continuous targets through the regression DLN [4]. This extension integrates a weighted sum of the final layer's logical rule outputs to produce a numerical prediction, successfully adapting the architecture to regression while maintaining its inherent interpretability. Despite the success of DLNs in tabular classification and their extension to regression, their application to time series data, particularly for classification tasks, remains largely unexplored.

Training AI models to understand images without costly labeled data remains a challenge. We combine two techniques--DINO (teacher-student learning) and Barlow Twins (redundancy reduction)--to create a model that learns better with fewer labels and less compute. While both DINO and Barlow Twins have independently demonstrated strong performance in self-supervised learning, each comes with limitations--DINO may be sensitive to certain augmentations, and Barlow Twins often requires batch sizes too large to fit on consumer hardware. By combining the redundancy-reduction objective of Barlow Twins with the self-distillation strategy of DINO, we aim to leverage their complementary strengths. We train a hybrid model on the MS COCO dataset using only 10\% of labeled data for linear probing, and evaluate its performance against standalone DINO and Barlow Twins implementations. Preliminary results show that the combined approach achieves comparable loss and classification accuracy to DINO while maintaining strong feature representations. Attention visualizations further suggest improved semantic segmentation capability in the hybrid model. This combined method offers a scalable, label-efficient alternative for training ViTs in resource-constrained environments.

Classification in the context of multi-label data streams represents a challenge that has attracted significant attention due to its high real-world applicability. However, this task faces problems inherent to dynamic environments, such as the continuous arrival of data at high speed and volume, changes in the data distribution (concept drift), the emergence of new labels (concept evolution), and the latency in the arrival of ground truth labels. This systematic literature review presents an in-depth analysis of multi-label data stream classification proposals. We characterize the latest methods in the literature, providing a comprehensive overview, building a thorough hierarchy, and discussing how the proposals approach each problem. Furthermore, we discuss the adopted evaluation strategies and analyze the methods' asymptotic complexity and resource consumption. Finally, we identify the main gaps and offer recommendations for future research directions in the field.

KL-Divergence Regularization Loss (KLL): We introduce a regularization loss function based on KL-divergence that explicitly measures and minimizes the probabilistic divergence between local and personalized model distributions. This theoretically grounded approach effectively prevents model drift while preserving the statistical characteristics of distributed client data, overcoming the limitations of conventional cosine similarity metrics. KL-Divergence-Prune Weighted Aggregation (KLPW A): We introduce a novel aggregation strategy that integrates KL-divergence-based distributional similarity, KL-Divergence-aggregation Weight (KLA W), and client-specific pruning ratios, Pruning-ratio-aggregation Weight (PRA W), into a unified weighting mechanism. This approach dynamically prioritizes clients that exhibit stronger alignment with the global model while contributing compact, efficiently pruned models. By jointly considering statistical consistency and model sparsity, KLPW A surpasses traditional aggregation methods in handling non-IID data distributions and substantially reduces communication costs. Sparse Activation Skipping (SAS): We present a mechanism of skipping pruned parameters during aggregation to enable the global model to be updated only with essential information. Cross-Round Recovery (CRR): To mitigate severe accuracy degradation caused by model pruning, we introduce the CRR, a two-stage pruning strategy. The CRR enables more precise decisions on whether to perform pruning, thus maintaining model accuracy after pruning. The remainder of this paper is organized as follows: Section II introduces related works and background of unsupervised federated person ReID.

Point cloud registration based on correspondences computes the rigid transformation that maximizes the number of inliers constrained within the noise threshold. Current state-of-the-art (SOTA) methods employing spatial compatibility graphs or branch-and-bound (BnB) search mainly focus on registration under high outlier ratios. However, graph-based methods require at least quadratic space and time complexity for graph construction, while multi-stage BnB search methods often suffer from inaccuracy due to local optima between decomposed stages. This paper proposes a geometric maximum overlapping registration framework via rotation-only BnB search. The rigid transformation is decomposed using Chasles' theorem into a translation along rotation axis and a 2D rigid transformation. The optimal rotation axis and angle are searched via BnB, with residual parameters formulated as range maximum query (RMQ) problems. Firstly, the top-k candidate rotation axes are searched within a hemisphere parameterized by cube mapping, and the translation along each axis is estimated through interval stabbing of the correspondences projected onto that axis. Secondly, the 2D registration is relaxed to 1D rotation angle search with 2D RMQ of geometric overlapping for axis-aligned rectangles, which is solved deterministically in polynomial time using sweep line algorithm with segment tree. Experimental results on 3DMatch, 3DLoMatch, and KITTI datasets demonstrate superior accuracy and efficiency over SOTA methods, while the time complexity is polynomial and the space complexity increases linearly with the number of points, even in the worst case.